Correspondence

Dietary Fat and the Risk of Breast Cancer

N Engl J Med 1996; 334:1606-1607June 13, 1996

Article

To the Editor:

The article by Hunter et al. (Feb. 8 issue)1 wrongly implies that lowering the total intake of fat in midlife does not reduce the risk of breast cancer.

They reanalyze seven prospective studies in which the dietary fat intake was not low enough to show an impact on breast cancer according to current data. Their reference value for fat intake was 30 to 35 percent of calories rather than less than 20 percent, which is supported as the level necessary for the prevention of breast cancer in post-menopausal women. The single dietary assessment at base line used in their analysis is not adequate to determine the role of dietary fat in breast cancer. People are notorious for fat intake, largely due to difficulty in estimating portion sizes.2 Underreporting is well documented and has been found to occur in up to 47 percent of cases.3 To reduce the incidence of underreporting, the study subjects must be taught how to recall accurately what they ate,2 and more thorough dietary assessments are necessary. Analysis of dietary-intake methods has shown telephone interviews to elicit the most reliable recall of food intake,4 and these could be used in large epidemiologic studies.

The small number of women in the analysis who met the criteria of having postmenopausal breast cancer and following a low-fat diet was an inadequate basis for a definitive conclusion. Therefore, the conclusion reached by the authors was not in the public interest. New techniques such as measurement of total energy expenditure with doubly labeled water5 and the Department of Agriculture Healthy Eating Index6 will help improve the accuracy of dietary assessment.

George L. Blackburn, M.D., Ph.D.
Masahiko Ishikawa, M.D.
Julie A. Smith, R.D.
Margaret Flynn, M.D.
Deaconess Hospital, Boston, MA 02215

6 References

1. 1

   Hunter DJ, Spiegelman D, Adami H-O, et al. Cohort studies of fat intake and the risk of breast cancer -- a pooled analysis. N Engl J Med 1996;334:356-361
   Full Text | Web of Science | Medline

2. 2

   Bolland JE, Yuhas JA, Bolland TW. Estimation of food portion sizes: effectiveness of training. J Am Diet Assoc 1988;88:817-821
   Web of Science | Medline

3. 3

   Lichtman SW, Pisarska K, Berman ER, et al. Discrepancy between self-reported and actual caloric intake and exercise in obese subjects. N Engl J Med 1992;327:1893-1898
   Full Text | Web of Science | Medline

4. 4

   Buzzard IM, Faucett CL, Jeffery RW, et al. Monitoring dietary change in a low fat diet intervention study: advantages of using 24-hour dietary recalls vs food records. J Am Diet Assoc (in press).
   

5. 5

   Beaton GH. Approaches to analysis of dietary data: relationship between planned analyses and choice of methodology. Am J Clin Nutr 1994;59:Suppl:253S-261S
   Web of Science | Medline

6. 6

   Kennedy ET, Ohls J, Carlson S, Fleming K. The Healthy Eating Index: design and applications. J Am Diet Assoc 1995;95:1103-1108
   CrossRef | Web of Science | Medline

To the Editor:

In their cohort studies, Hunter et al. concluded that there is no evidence of a positive association between fat intake and the risk of breast cancer in the context of a Western lifestyle. However, in rural Africa, where breast cancer is very uncommon, a low intake of fat is only one factor. In South Africa, Africans still live relatively traditionally in many rural areas. In three isolated rural hospitals responsible for the needs of about 400,000 people, among about 16,000 admissions annually there were approximately 8 cases of breast cancer (although there were plenty of cases of cervical, esophageal, and liver cancer), for an incidence of 4 per 100,000 women, 1/10 of the incidence among U.S. women.1 The onset of menstruation among African girls occurs an average of 1 to 1 1/2 years later than among white girls, menstrual cycles are less regular, about 30 percent become pregnant as teenagers, and parity is high, with long intervals of breast-feeding and associated amenorrhea.2 In brief, ovulatory cycles are far fewer in African than in white women. Energy intake is relatively low, fat intake supplies about 20 percent of energy, and the average daily fiber intake is about 25 to 35 g. The level of physical activity is high. Probably, a low or reduced fat intake has limited protective capacity against breast cancer unless it is associated with some of the other protective factors listed.

Hunter et al. suggest that aspects of diet during childhood and adolescence may be associated with the risk of breast cancer decades later. A recent local study revealed that the mean daily energy intake of five-year-old rural African girls was about 1000 kcal, as compared with nearly double that amount in corresponding white girls. The mean respective daily total fat intakes were 27 g and 77 g.3

Should the prevention of breast cancer depend on a person's compliance with a number of the protective factors listed, then primary prevention of breast cancer in Western and some urban developing populations would seem to be almost impossible, even in those with a familial risk.

Alexander R.P. Walker, D.Sc.
Betty F. Walker, Dipl.Dom.Sci.
South African Institute for Medical Research, Johannesburg 2000, South Africa

3 References

1. 1

   Whelan SL, Parkin DM, Masuyer R, World Health Organization. Patterns of cancer in five continents. Lyon, France: International Agency for Research on Cancer, 1990. (IARC scientific publications no. 102.)
   

2. 2

   Walker AR, Walker BF, Stelma S. Is breast cancer avoidable? Could dietary changes help? Int J Food Sci Nutr 1995;46:373-381
   CrossRef | Web of Science | Medline

3. 3

   Cleaton-Jones P, Granath L, Richardson B, eds. Dental caries, nutrient intake, dietary habits, anthropometric status, oral hygiene, and salivary factors and microbiota in South African black, Indian and white 4-5 year-old children. Parow: South African Medical Research Council, 1991.
   

Author/Editor Response

The authors reply:

To the Editor: The points raised by Blackburn et al. relate to the possibility of error in base-line dietary assessment in the seven prospective studies that we analyzed. Precisely because of this concern we limited our analyses to studies in which the error in estimation of usual dietary fat intake, as measured by food-frequency questionnaires, could be assessed by comparison with fat intake measured by a reference method (either multiple food records or multiple 24-hour periods of food-intake recall). The average correlation coefficients were approximately 0.5, suggesting reasonably accurate reporting. With these data, we were able to correct the observed relative risk and confidence intervals for bias due to this degree of measurement error,1 and we still did not observe any substantial positive association between total fat intake and the risk of breast cancer. Blackburn et al. suggest telephone interviews as a method of obtaining reliable dietary information; it is unlikely that this could be used in large epidemiologic studies, since many days of recall of food intake are necessary for each subject to approximate the average intake of most nutrients, because of the high day-to-day variability. Similarly, although measurements of energy expenditure using doubly labeled water and the Department of Agriculture Healthy Eating Index may have their place, it is hard to see how they can really be helpful in very large population-based cohort studies.

Blackburn and colleagues are correct that only a modest number of subjects reported consuming less than 20 percent of calories from fat. The fact that this number is small is expressed in the 95 percent confidence interval for this estimate; the lower bound (0.83) rules out a substantial reduction in risk. Furthermore, the significant increase in risk we observed among women with less than 15 percent of energy intake from fat is hard to reconcile with a substantial reduction in risk.

Walker and Walker suggest multiple reasons why breast-cancer rates are so low in South Africa. With so many differences in reproductive risk factors, exercise, childhood diet, and energy intake, it is clear that choosing fat intake as the sole explanation for the international variation in breast-cancer rates is scientifically unsound. We agree that these data from South Africa suggest that multiple factors are almost certainly responsible and that aspects of childhood diet and growth are worth taking seriously in the search for the causes of breast cancer.

David J. Hunter, M.B., B.S.
Donna Spiegelman, Sc.D.
Walter Willett, M.D.
Harvard School of Public Health, Boston, MA 02115

1 References

1. 1

   Rosner B, Spiegelman D, Willett WC. Correction of logistic regression relative risk estimates and confidence intervals for measurement error: the case of multiple covariates measured with error. Am J Epidemiol 1990;132:734-745
   Web of Science | Medline

Citing Articles (4)

Citing Articles

1. 1

   Marion M. Lee, Scarlett S. Lin. (2000) D IETARY F AT AND B REAST C ANCER. Annual Review of Nutrition 20:1, 221-248
   CrossRef

2. 2

   ERNST L WYNDER, LEONARD A. COHEN, BARBARA L. WINTERS. (1997) The Challenges of Assessing Fat Intake in Cancer Research Investigations. Journal of the American Dietetic Association 97:7, S5-S8
   CrossRef

3. 3

   Masakuni Noguchi, Takao Taniya, Takeo Kumaki, Nagayoshi Ohta, Hirohisa Kitagawa, Kazuo Kinoshita, Mitsuharu Earashi, Ryo Yagasaki, Masahide Minami, Futoshi Kawahara, Hiroshi Tsuyama, Koichi Miwa. (1997) Dietary fat and breast cancer: A controversial issue. Breast Cancer 4:2, 67-75
   CrossRef

4. 4

   ARP Walker, I Segal. (1996) Fibre and colorectal cancer. The Lancet 348:9032, 956-957
   CrossRef